AIM: To summarize the current knowledge about the potential relationship between hepatitis C virus (HCV) infection and the risk of several extra-liver cancers.

METHODS: We performed a systematic review of the literature, according to the Preferred Reporting Items for Systematic reviews and Meta-Analysis (PRISMA) Statement. We extracted the pertinent articles, published in MEDLINE and the Cochrane Library, using the following search terms: neoplasm/cancer/malignancy/tumor/carcinoma/adeno-carcinoma and non-Hodgkin lymphomas, kidney/renal-, cholangio-, pancreatic-, thyroid-, breast-,oral-, skin-, prostate-, lung-, colon-, stomach-, haematologic. Case series, case-series with control-group, case-control, cohort-studies as well as meta-analyses, written in English were collected. Some of the main characteristics of retrieved trials, which were designed to investigate the prevalence of HCV infection in each type of the above-mentioned human malignancies were summarised. A main table was defined and included a short description in the text for each of these tumours, whether at least five studies about a specific neoplasm, meeting inclusion criteria, were available in literature. According to these criteria, we created the following sections and the corresponding tables and we indicated the number of included or excluded articles, as well as of meta-analyses and reviews: (1) HCV and haematopoietic malignancies; (2) HCV and cholangiocarcinoma; (3) HCV and pancreatic cancer; (4) HCV and breast cancer; (5) HCV and kidney cancer; (6) HCV and skin or oral cancer; and (7) HCV and thyroid cancer.

RESULTS: According to available data, a clear correlation between regions of HCV prevalence and risk of extra-liver cancers has emerged only for a very small group of types and histological subtypes of malignancies. In particular, HCV infection has been associated with: (1) a higher incidence of some B-cell Non-Hodgkin-Lymphoma types, in countries, where an elevated prevalence of this pathogen is detectable, accounting to a percentage of about 10%; (2) an increased risk of intra-hepatic cholangiocarcinoma; and (3) a correlation between HCV prevalence and pancreatic cancer (PAC) incidence.

CONCLUSION: To date no definitive conclusions may be obtained from the analysis of relationship between HCV and extra-hepatic cancers. Further studies, recruiting an adequate number of patients are required to confirm or deny this association.

Hepatitis C virus (HCV) is a major global health problem, because it represents a very important cause of mortality, morbidity and resource utilization. Although remarkable differences are detectable in the world, depending on geographical areas and ethnicity, it is estimated that the prevalence of HCV infection is about 2% worldwide (Figure 1)[1]. Approximately 180 million people carriers this pathogen persistently[2]. HCV chronic infection can lead to a necro-inflammatory liver disease, with different pattern of severity and course. This condition is associated with an increased risk of cirrhosis, liver failure and hepatocellular carcinoma[3]. Although liver is the main target for HCV, it is now well-known that this pathogen may induce extra-hepatic pathological conditions, including mixed cryoglobulinemia, porphyria cutanea tarda, membranoproliferative glomerulonephritis, Sjögren’s syndrome, thyroiditis, a high prevalence of autoantibodies[4] as well as Central and Peripheral Nervous System demyelinating disorders[5]. Several of these manifestations are thought to be caused by the host immune response to this micro-organism and not by a direct viral cytopathic effect. In particular, chronic antigenic stimulation by HCV promotes B-lymphocyte clonal expansion, with the production and release of monoclonal and polyclonal antibodies and generation of immune complexes[6]. Their deposition in small vessels and glomerular capillary walls induces complement activation and, as consequence, tissue injury[7]. In addition, several studies have shown that HCV may infect organs and tissues other than the liver. In particular, presence of antigens, genome and/or replicative sequences of HCV have been detected in several extra-hepatic localizations, such as peripheral blood cells (i.e., neutrophils, T- and B-lymphocytes)[8,9] or kidney[10], skin[11,12], oral mucosa[13], salivary glands[14] and pancreas tissues as well as, in a small number of cases, from heart, gallbladder, intestine and adrenal glands tissues[15,16].

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Prevalence of chronic hepatitis C virus status in different coutries worldwide. HCV: Hepatitis C virus.

Although HCV antigens and replicative forms have been detected in various extra-hepatic sites, the possible role on the onset of malignancies in these organs is still under investigation. Some evidences have recently suggested the possibility that this pathogen may be associated with the development of a wide spectrum of hematologic or solid cancers, such as non-Hodgkin lymphomas, biliary duct-, bladder-, renal-, pancreatic-, thyroid-, breast- and prostate-carcinomas. Here we summarize the current knowledge about the potential link between HCV infection and risk of these malignancies and we performed a systematic review of the literature that reports the prevalence of HCV infection in patients, suffering from above mentioned malignancies.

The search of MEDLINE and Cochrane Library produced the following citations: (1) haematopoietic malignancies: 1424; (2) biliary ducts-/cholangio-: 616; (3) renal-/kidney-: 891; (4) pancreatic-: 244; (5) thyroid-: 126; (6) breast-: 180; (7) lung-: 247; (8) stomach-: 141; (9) colon-: 115; (10) skin-/oral-: 598; (11) bladder-: 150; and (12) prostate-carcinomas: 43.

After a preliminary review of the titles and/or abstracts with the exclusion of non-pertinent articles, we obtained these results: (1) haematopoietic malignancies, including lymphomas/non-Hodgkin lymphomas: 126; (2) biliary ducts-/cholangio-: 48; (3) renal-/kidney-: 10; (4) pancreatic-: 15; (5) thyroid-: 11; (6) breast-: 8; (7) lung-: 3; (8) stomach-: 2; (9) colon-: 5; (10) skin-/oral-: 11; (11) bladder-: 3; and (12) prostate-carcinomas: 5.

We screened the potentially relevant studies and, in accordance with predefined criteria, we identified and considered in our systematic review the following number of studies: (1) haematopoietic malignancies: 108 articles considered, 6 reviews/meta-analyses, 12 papers excluded[17-146]; (2) biliary ducts/cholangiocarcinoma: 36 articles considered, 3 reviews/meta-analyses, 9 papers excluded[147-195]; (3) renal/kidney: 8 articles considered, 2 papers excluded[118,146,196-203]; (4) pancreatic: 9 articles considered, 3 reviews/meta-analyses, 3 papers excluded[118,146,170,197,204-214]; (5) thyroid: 7 articles considered, 4 papers excluded[52,73,118,146,196,197,215-217]; (6) breast: 6 articles considered, 2 papers excluded[117,118,146,196,197,202,218,219]; (7) lung: 2 articles meeting inclusion criteria, 1 not[146,196,197]; (8) stomach: 2 articles meeting inclusion criteria[146,220]; (9) colon: 3 articles meeting inclusion criteria, 2 not[118,146,196,197,202]; (10) skin/oral: 10 articles considered, 1 paper excluded[118,146,197,221-228]; (11) bladder: 3 articles meeting inclusion criteria[118,146,197]; and (12) prostate-carcinomas: 4 articles meeting inclusion criteria, 1 not[118,146,196,197,202] (Tables 1, 2, 3, 4, 5, 6 and 7). A limited number of identified studies were formally designed as “cohort-“ or “case-control” trials, adequately reporting inclusion criteria for the control group, such as “odds ratios” after adjustment for the most important confounding factors, or showing that cases and controls were matched by sex and age. Whether these data were not indicated, but an acceptable series of healthy subjects or patients with different diseases were recruited for comparison and were described, we defined the considered study, as: “case series with control group”.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Number of studies and meta-analyses, assessing the association between hepatitis C virus infection and different types of malignancies, included in the present systematic review. References are reported in the supplementary section.

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Age-standardized incidence rates of each malignancy per 100000 person-years are reported for sex and for different countries. A: Incidence rates of non-hodgkin-lymphomas (GLOBOCAN 2012); B: Incidence rates of cholangiocarcinoma (by Shin HR, Asian Pacific Journal of Cancer Prevention 2010; 11); C: Incidence rates of pancreatic cancer (GLOBOCAN 2012); D: Incidence rates of breast cancer (GLOBOCAN 2012); E: incidence rates of kidney cancer (GLOBOCAN 2012); F: Incidence rates of oral/skin cancers (GLOBOCAN 2012); G: Incidence rates of thyroid cancer (GLOBOCAN 2012).

On the basis of our results, we summarised some of the main characteristics of retrieved trials, which were designed to investigate the prevalence of HCV infection in each type of the above- mentioned malignancies. In particular, we created a main table and included a short description in the text for each of these tumours, whether at least five studies, evaluating this parameter and meeting inclusion criteria, were available in literature. In each of these tables, we reported the following data of studies considered: first author’s name, study design, year of publication, country of origin, matching criteria, number of cases and controls, diagnostic methods to detect each malignancy, percentage of HCV-positive cases with 95% confidence intervals (CIs) and main conclusions. CIs for each proportion were calculated according to normal distribution or binomial distribution as appropriate. Accordingly to these pre-definite criteria, we created the following sections and the corresponding Tables: (1) HCV and haematopoietic malignancies (Table 1); (2) HCV and cholangiocarcinoma (Table 2); (3) HCV and pancreatic cancer (Table 3); (4) HCV and breast cancer (Table 4); (5) HCV and kidney cancer (Table 5); (6) HCV and skin or oral cancer (Table 6); and (7) HCV and thyroid cancer (Table 7). Furthermore, we created additional tables, reporting both the studies not considered in our systematic review and the meta-analyses, assessing the association between HCV infection and risk of each human malignancy (see Tables 1-7). A summary of number of studies and meta-analyses is shown in Figure 2. Age-standardized incidence rates of each malignancy per 100000 person-years are reported for sex and for different countries in Figure 3[229].

To our knowledge, this is the first study aimed to review systematically the prevalence of HCV infection in a wide spectrum of human malignancies, to summarise the retrieved data and to discuss the possible role of this pathogen in the genesis of the discussed tumours. Since several years the possible involvement of different viruses in human carcinogenesis has been reported, with increasing frequency, in a large series of epidemiological studies. Recently, the International Agency for Research on Cancer (IARC) has comprehensively assessed and confirmed the human carcinogenicity of 7 viral agents, including HCV. In particular, it has been recognised that HCV acts as an indirect carcinogen, by promoting and maintaining a state of chronic inflammation in infected sites[248]. This event is now a well-known condition that is involved in the process of hepatic carcinogenesis. However, even if liver is the main target of HCV, its tropism for this organ is not exclusive. In particular, viral antigens and genomes have been also detected in extra-hepatic tissues[249]. All these evidences have contributed to consider the possible role of this pathogen as a pro-cancerous agent also in organs other from liver. On the basis of this assumption, as it has been observed for the strong relationship between HCV and hepatocellular carcinoma (HCC) development, it is conceivable to think that a similar ecological correlation might exist between HCV infection and a major risk of some extra-hepatic cancers in regions where the prevalence of this pathogen is high in comparison with geographical areas with a lower one. However, to date, the possibility that this virus may be involved in the carcinogenesis of organs other than liver has been systematically investigated only for a very limited number of malignancies. In particular, available studies have been mainly focused on hematopoietic malignancies and on cholangiocarcinomas, also on the basis of some epidemiological observations, reporting a major risk of mixed cryoglobulinemia and monoclonal gammopathy in HCV positive patients[250] as well as of a higher incidence of cholangiocarcinoma and other types of human cancer other than liver in subjects with cirrhosis, irrespective of etiology[251]. In this study, enrolled patients suffered from alcoholic-, primary biliary-, and chronic-hepatitis-related cirrhosis, whereas in a group of patients the causes of this pathological condition were non-specified. Unfortunately, no detailed information was available, concerning the HCV status in the cohort of subjects with persistent viral hepatitis. Even if this observation was very interesting, it has not stimulated the achievement of studies investigating specifically the possible association between hepatitis viruses (in particular HCV) and human cancers other than HCC. Only in the last years this idea has gained interest an increasing number of trials have been designed and carried out with the purpose. Nevertheless, to date, few data are available yet and no final or univocal conclusions may be drawn. However, putting together the results of published studies on the possible association between HCV and risk of NHLs, cholangio-, pancreatic-, breast-, renal-, skin/oral- and thyroid-cancers, the reports concerning the estimates of prevalence of HCV infection worldwide as well as the tables on the age-standardized incidence rates of the mentioned malignancies per 100000 person-years in different geographical areas, some interesting consideration may be made (Figures 2 and 3). Actually, a clear correlation between regions of HCV prevalence and risk of extra-liver cancers has emerged only for a very small group of types and histological subtypes of malignancies. In particular, HCV infection has resulted to be associated with: (1) a higher incidence of some B-cell NHLs types, in countries, where an elevated prevalence of this pathogen is detectable, accounting to a percentage of about 10%. Furthermore, an additional factor potentially confirming the causal role between HCV and lymphomas, in particular B-cell NHLs, is represented by the observation of the regression of a significant percentage of low-grade B-cell NHLs after HCV eradication by means of an efficacious antiviral treatment. Early evidences, concerning this assumption, date back to the end of Twentieth Century with anecdotal reports and the beginning of the following Century, when further studies, enrolling a wider number of patients, were carried out for the first time[252]; and (2) an increased risk of intra-hepatic cholangiocarcinoma development in subjects with anti-HCV positivity, as reported by a large number of available studies with OR ranging from 3.42 to 4.84. According to an epidemiological view-point, it has to be considered the following evidence: liver flukes are associated with an increased risk both for ICC and ECC, but in endemic-nations (such as Taiwan, Singapore and Korea) and in other Eastern areas with low rates of this type of infection, the incidence of ICC is more elevated in comparison with ECC. On the other hand, in the western areas ECC incidence is higher than that of ICC. Taking into account HCV epidemiology, the prevalence of this virus presents a wide variability worldwide. More elevated percentages are detectable in developing areas, such as in South-Eastern-Asia and Egypt, intermediate ones in Southern Europe and the lowest ones in Northern Europe and America. Therefore, a correlation between HCV prevalence and ICC incidence seems to emerge from these observations.

Concerning the possible association between the infection caused by this hepato-tropic virus, surprisingly, although a higher risk of PADC in HCV-positive patients has been observed in some trials and it has been reported in the available meta-analyses, age-standardized rates of this cancer present interesting geographical variations. In particular, PADC incidence worldwide is 3-4 times higher in more economically developed countries as well as in northern area of the world, where HCV prevalence is lower in comparison with less developed countries. Different reasons may explain these results, including the accuracy of diagnostic methods used to diagnose pancreatic malignancies in distinct geographical regions worldwide and of data to assess the rates of incidence of PDAC. However, it is well-known that on the basis of a morphogenetic viewpoint, pancreas and liver share several characteristics in their embryological development, arising from common multi-potent cells of endoderm origin. Therefore, HCV might replicate also in pancreatic cells as it does in hepatocytes[253,254]. Furthermore, according to epidemiological studies, type 2 diabetes represents a risk factor for PAC[255,256] with chronic hyperglycemia and hyperinsulinemia as proposed pathogenetic mechanisms involved in the promotion of this type of malignancy. Both conditions may induce proliferation, decrease apoptosis and promote invasion ability of pancreatic cancer cells[257-259]. A recent and interesting systematic review and dose-response meta-analysis, assessing blood glucose concentration and risk of pancreatic cancer, has shown that every 0.56 mmol/L (10 mg/dL) increase in fasting blood glucose , a 14% enhancement in the rate of PADC occurs[260]. On the other hand, it is now accepted that HCV infection is associated with an increased risk of type 2 diabetes[261,262]. Therefore, HCV infection, promoting and diabetes might act in cooperation with hyperinsulinemia, hyperglycemia in the promotion of PAC. This consideration may contribute to explain the epidemiological evidence of a more elevated incidence of this cancer in the most economically developed countries, where high rates of obesity, metabolic syndrome and diabetes are observed, in comparison with less developed regions in the world.

On the other hand, up to now no definitive and univocal conclusions may be obtained from the analysis of relationship between HCV and breast-, renal-, skin/oral- and thyroid-cancers. Most of the available studies, in particular a large part of cohort trials, have not confirmed the existence of these associations. However, the lack of similar evidences may be only apparent and some elements potentially limiting the conclusions of these reports have to be taken into account. In particular, most of available data have been obtained from some retrospective cohort-trials. Although the large size and the lengthy follow-up of this type of research provides the statistical power to obtain adequate information on cancer risk in the investigated population, the retrospective nature of these studies has to be considered. The use of routinely collected administrative data, based on population registries may represent a potential limiting factor in these trials. Possible errors in diagnosis, in codes of cancers or infective diseases and in reported data may affect the hospitalization records in some case/control subjects as well as the possibility that in several countries worldwide only a part of general population is included in national cancer registry may influences the obtained results. In addition, some of these trials have evaluated the incidence of a wide spectrum of malignancies and they have been not designed to assess a specific type of human tumors. It has to be also underlined that, according to age-standardized incidence rates of these neoplasms, remarkable temporal changes in human cancer trends have been observed worldwide. The combined analysis of these figures, as performed some years ago for HCV-related liver cancer, has induced some authors to hypothesize a possible role of HCV in the development of these malignancies. For example, in Japan, Tanaka et al have examined temporal trends for HCC incidence rates in a period ranging from 1981 and 2003 in Osaka Prefecture. They provided an explanation in the context of HCV infection rates. According to these findings, they noted that in that span of time the incidence peak of HCC was detectable in men during their 50s, 60s and 70s of age in 1986, 1995 and 2000 and then it was progressively decreasing. He postulated that the observed trend was due to the restriction of virus transmission[263]. As previously suggested, it should be taken into account that an enhanced incidence of ICC has been observed in the most developed regions in the years, ranging from 1980 to 2000, has been considered as caused by the parallel increase of HCV infection in these areas. This hypothesis has been definitively shown in United States[264]. In addition, the relationship between HCV and risk of ICC has been assessed in patients with different degree of hepatic damage. It has been demonstrated that the probability of ICC development increases as the hepatic damage impairs[167]. Furthermore, a previous Italian research had suggested that HCV might have a role in thyroid cancer. This represents a rather rare malignancy, but its standardized incidence ratio has progressively increased between the end of ‘80s and the beginning of ‘90s in several well-developed countries, including Italy. In this country, in the same period of time an increasing prevalence of HCV infection was observed[247]. Unfortunately, no additional trials have been performed, with the purpose to assess the possible impact of HCV infection worldwide on the age-standardized incidence rates of the aforementioned malignancies in different geographical areas and to distinguish the potential contribution of this pathogen as pro-carcinogenic agent from other risk factors. Therefore this field of research still remains largely unexplored. In addition, it has to be taken into account that HCV represents a common cause of cirrhosis development and this condition itself is a well-known independent risk factor for the development of a wide spectrum of human malignancies different from liver. Several mechanisms may be responsible for carcinogenetic role of this pathogen. It may act indirectly as well as directly. Both these activities have been proposed for this virus and plausible evidences reported in literature. To date the existence of an indirect action of this microorganism is the most convincing pathogenic mechanism. In particular, as widely reported for liver, HCV might promote in extra-hepatic organs a persistent inflammation and induce the cancerous transformation, as a consequence of a progressive rearrangement of their structure. This event might play a role in the development of all the types of the aforementioned malignancies and not only involved in hepatic carcinogenesis. This process is characterized by the interaction and the cooperation among viral-related proteins, homing-cells (specific-tissue-cells, depending on the considered organ, endothelial cells, and stem cells), not homing cells (lymphomononuclear and polymorfonuclear specific and nonspecific cells), cytokines, costimulatory molecules and additional biological mediators (i.e., PGs and oxidants). This situation promotes a self-maintaining and amplifying loop, in which HCV stimulate, in turn, PGE2, enzymes (such as cyclo-oxygenase-2 or COX-2), growth factors, interleukins production and cellular signaling pathway function. The complex cooperation and interaction among these mediators is one of the main factors responsible for final outcome: viral control with recovery or its persistence in the infected-organs, with progressive development of a tissue necro-inflammatory process, potentially evolving toward malignant transformation. Presence of inflammation favours genetic instability in cells and increases the probability which further genetic and epigenetic alterations arise. Perturbance of homeostasis in adult-cells may re-modulate activity and expression of genes as well as of transcription factors that govern their regeneration and/or differentiation programs as well as the processes involved in energy production. In addition, alteration of function of some intracellular pathways, such as K-RAS, Hedgehog-, Jak-STAT-, Notch-and TGF-β signalling-cascades, may also play a key role in carcinogenesis. In particular, this events may be induced not only by inflammation itself, but also by the direct action of some viral proteins such as: core- and NS5A on the intracellular cascades pathways function. These viral elements may affect the levels of activities of the afore-mentioned signalling-cascades and contribute to deregulation of cell-cycle checkpoint controls. To date a few studies have been performed with the aim to specifically investigate the pathogenetic mechanisms, potentially involved in the development of extra-liver malignancies in anti-HCV positive patients. The majority of reports upon this topic concerns lymphoproliferative malignancies. A study has described an increased expression of genes associated with lymphomagenesis in peripheral blood B cells of chronic HCV positive patients[265].

Furthermore, it has been shown that the telomere deletion of 1p36.3 in B-cell NHLs is significantly more frequent in patients with HCV infection in comparison with anti-HCV negative individuals[266]. Deletion of the 1p36 genomic locus is associated with the loss of p73, a tumour suppressor gene, that may be inactivated both in lymphomas and in other human cancers[267,268]. On the basis of available studies, this topic is progressively acquiring an increasing importance, but to date only some definitive conclusions have been obtained, while a large number of questions still remain unanswered. Further well-designed trials, enrolling an adequate number of patients as well as focusing on populations of different geographical areas and involving larger series of patients are required to confirm or deny this association as well as to identify the pathogenetic mechanisms, potentially involved in HCV-associated human carcinogenesis.

The authors thank Dr. Simonetta Righi, Biblioteca Centralizzata, Policlinico S Orsola-Malpighi, Università di Bologna, Bologna, Italy, for her support in the search of scientific bibliography and Mr. Fabio Castellini, Dipartimento di Emergenza-Urgenza, Ospedale di Budrio, Budrio, Bologna, Italy, for his active support in clinical research.